In at least three XMM-Newton spectra of X-ray GRB afterglows, the
presence of a thermal component has been claimed (R02; Watson et al. 2002). We have studied the geometrical conditions that the thermal
material must satisfy in order to contribute significantly to the
early X-ray afterglow. We first concluded that the time scale during
which the emission is observed must be set by the light crossing time
of the emitting region. This implies that standard free-free equations
that relate the emission integral to the luminosity cannot be used and
a more general equation (Eq. (1)) has to be adopted. The
implications of this are of great importance. In fact, the use of the
standard formalism (Eq. (2)) to compute the EI led R02 to
underestimate the emission integral for GRB 011211 by a factor
104, and conclude that thermal emission from a shell with
cm can be a self consistent solution. We showed that,
with the correct treatment of light-crossing effects, this is not the
case. We then studied the possibility of (i) flash heating and (ii)
steady heating. In both cases, we conclude that the thermal material
must be extremely clumped in order to contribute significantly to a
typical X-ray afterglow. In the steady heating case, the thermal
material is continuously heated for a time longer than the cooling
one, reaching a stationary state (albeit for a time smaller than the
light-crossing one). We showed that in this case the material has to
be so extremely clumped that the clumps are dissolved by the increased
temperature in a very short time scale, so that an additional source
of confinement must be envisaged. In the case of flash heating, a self
consistent solution can be found, requiring a less extreme (even if
yet compelling) clumping. In this case, the thermal material must be
located far from the source and therefore, in order to preserve the
fast time variability, only a small portion of it has to be
heated. Assuming that the same density is spread all over the GRB
explosion site, we derive that the total mass involved is large (
), larger than even a type-II supernova remnant. A
moderate asymmetry of the remnant could however ameliorate this
requirement. It is interesting to note that, even though such
clumpy and dense environments seem extreme, they were already inferred
independently to account for the lack of ionization features in
GRB 000210 (Piro et al. 2002) and to explain the high energy emission
in GRB 940217 (Katz 1994).
To date there is not yet compelling evidence that there are indeed thermal components in the early X-ray afterglow of GRBs. It is however intriguing then in three out of four afterglows observed by XMM-Newton a thermal model yields a better fit than an absorbed power-law (R02, Watson et al. 2002a). In the case of GRB 011211 (R02) it is particularly difficult to explain the absence of an iron or nickel line, while reflection models seem to give a better explanation of the observed line ratios (Lazzati et al. 2002b).
Nonetheless, should a thermal component be confirmed in the X-ray afterglow of at least a sub-fraction of GRBs, it would be difficult to avoid the conclusion that (i) the GRB explosion site is surrounded by a massive shell of matter, similar to a 1-2 year old supernova remnant (as expected in some versions of the Supranova model, Vietri & Stella 1998); (ii) that this remnant is extremely clumped, suggesting that some precursor activity has been taking place. Whether such a clumpiness can help make the prompt emission in an external shock scenario (Dermer et al. 2000) heavily depends on the presence of clumping at smaller scales, a test that cannot be done with thermal emission from the clumps, since it would contribute in any case a negligible flux to the early afterglow emission.
Acknowledgements
I am grateful to G. Ghisellini, P. Kumar, M. J. Rees and F. Tavecchio for useful comments and discussions. I acknowledge financial support from the PPARC.
Copyright ESO 2003